import gradio as gr
from transformers import pipeline

pipe = pipeline(
    task="text-generation",
    model="UpMath/Thai-HomeworkGen-v3",
    device_map="auto",
    trust_remote_code=True
)

def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    context = system_message + "\n"
    for user, assistant in history:
        if user:
            context += f"User: {user}\n"
        if assistant:
            context += f"Assistant: {assistant}\n"
    context += f"User: {message}\nAssistant:"

    outputs = pipe(
        context,
        max_new_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
        do_sample=True,
    )

    response = outputs[0]["generated_text"]
    
    response_only = response[len(context):].strip()
    yield response_only

demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
)

if __name__ == "__main__":
    demo.launch()